Monday, August 27, 2007

KL Day 2 - Faking up Is Hard to Do

Worst part of my job is having to do something that you have no idea about. Officially, I'm here in KL to do some MVI work - making sure our 3G SGSN works with a competitor's RNC. My colleague, whom I replaced, assured me that most of the work has been done, and all I need to do is babysit the systems and provide logs if necessary. Easy peasy - that I can do while enjoying my time in KL.

First day I walked into the customer site, the project manager informed me that the PS is not working. Two out of four ATM links are up, CS is working, but not PS, and I need to fix it quick. What?! This is not my idea of a junket. First thing to do is to bring up the other two links. Bounced the links - no. Recreated the entries - no. Modified a wrong configuration - no. Turns out the RNC guy didn't create the link entries on his end. All links up - yes. PS working - no. RANAP just won't come up. It was a pretty stressful situation, I tell you, with everyone standing around watching you struggle to make things work. I decide to use my "Call-A-Friend" lifeline. After extensive consultation with my colleagues, we still don't know what's the problem. One option is to restart the TU and SMM cards. If in doubt, reboot. Another option is to do an SMM debug trace. This is going to be hairy as this is a production system - it'll produce humongous amounts of logs and might even bring down the whole box. I decide to leave it till tomorrow.

Later tonight, I got an advice from another colleague to create a dummy RNC, apply-save, then delete the dummy RNC config. Said it worked for him before. That'll allow me to sleep a little bit better tonight.

No comments: